ORIGINAL ARTICLE

Reconstitution of huPBL-NSG Mice with Donor-MatchedDendritic Cells Enables Antigen-Specific T-cell Activation

Airi Harui & Sylvia M. Kiertscher & Michael D. Roth

Received: 1 April 2010 /Accepted: 18 May 2010 /Published online: 8 June 2010# The Author(s) 2010. This article is published with open access at Springerlink.com

Abstract Humanized mouse models provide a uniqueopportunity to study human immune cells in vivo, buttraditional models have been limited to the evaluation ofnon-specific T-cell interactions due to the absence ofantigen-presenting cells. In this study, immunodeficientNOD/SCID/IL2r-γnull (NSG) mice were engrafted withhuman peripheral blood lymphocytes alone or in combina-tion with donor-matched monocyte-derived dendritic cells(DC) to determine whether antigen-specific T-cell activa-tion could be reconstituted. Over a period of 3 weeks,transferred peripheral blood lymphocytes reconstituted thespleen and peripheral blood of recipient mice withpredominantly human CD45-positive lymphocytes. Ani-mals exhibited a relatively normal CD4/CD8 ratio (average1.63:1) as well as reconstitution of CD3/CD56 (averaging17.8%) and CD20 subsets (averaging 4.0%). Animals

reconstituted with donor-matched CD11c+ DC also dem-onstrated a CD11c+ population within their spleen, repre-senting 0.27% to 0.43% of the recovered human cells withconcurrent expression of HLA-DR, CD40, and CD86.When immunized with adenovirus, either as freereplication-incompetent vector (AdV) or as vector-transduced DC (DC/AdV), there was activation andexpansion of AdV-specific T-cells, an increase in Th1cytokines in serum, and skewing of T-cells toward aneffector/memory phenotype. T-cells recovered from animalschallenged with AdV in vivo proliferated and secreted aTh1-profile of cytokines in response to DC/AdV challengein vitro. Our results suggest that engrafting NSG mice witha combination of lymphocytes and donor-matched DC canreconstitute antigen responsiveness and allow the in vivoassessment of human immune response to viruses,vaccines, and other immune challenges.

Keywords dendritic cell . adenovirus . vaccination .

xenotransplant . cytokine . humanized mouse

Introduction

While inbred strains of conventional and geneticallymodified mice have provided an unparalleled opportunityto examine the process of antigen presentation and thecharacteristics of adaptive T-cell immunity, translating thesefindings into new approaches for human immunotherapyhas proven problematic (Mestas and Hughes 2004; Ju et al.2010). Responses to cytokine therapy, vaccines, genetherapy, adjuvants, and adoptive cell transfers that arereproducible and robust in the mouse are often rare whentranslated into human clinical trials. This is not surprisinggiven that human dendritic cells (DC) are distinctively

This study was presented at the 15th Scientific Conference of theSociety on NeuroImmune Pharmacology, Wuhan, China, 2009.

This study was supported by funding from the NIH/National Instituteon Drug Abuse (RO1 DA03018), the Department of Defense ProstateCancer Research Program, Stewart and Lynda Resnick, the ProstateCancer Foundation, the UCLA Department of Urology, and Coresupport from the Jonsson Comprehensive Cancer Center.

A. Harui : S. M. Kiertscher :M. D. RothDivision of Pulmonary & Critical Care Medicine,David Geffen School of Medicine at UCLA,Los Angeles, CA 90095-1690, USA

M. D. Roth (*)Division of Pulmonary & Critical Care Medicine,Department of Medicine, David Geffen School of Medicine atUCLA,Los Angeles, CA 90095-1690, USAe-mail: mroth@mednet.ucla.edu

J Neuroimmune Pharmacol (2011) 6:148–157DOI 10.1007/s11481-010-9223-x

different from those of the mouse, as are the antigens ofinterest, the major histocompatibility complex (MHC)repertoire, toll-like receptor expression, cytokine responses,and many other features. Genetic diversity is also a majorfactor, as are the susceptibility and prior exposure to manyhuman pathogens and immune challenges (Mestas andHughes 2004; Zaiss et al. 2009). Even transgenic animalsthat express a human MHC haplotype or spontaneouslydevelop organ-specific tumors have not proven to bepredictive of integrated human immune responses (Pascoloet al. 1997).

One approach for bridging the gap between mousemodels and human clinical trials has been to develophumanized mice—immunodeficient strains which permitthe engraftment and expansion of human immune cells invivo (Shultz et al. 2007; Pearson et al. 2008). A variety ofdifferent humanized mouse models have been developedand fall into two distinct categories: (1) those with a slowbut more extensive immune reconstitution resulting fromstem cell transfer (Lapidot et al. 1992; Greiner et al. 1998;Ito et al. 2002) or (2) those with rapid reconstitution basedon the transfer of fully differentiated human peripheralblood cells (Mosier et al. 1988; Hesselton et al. 1995; Mayet al. 2005). We have focused on the latter, not only for thespeed of the reconstitution but for the prospect that thereconstituted immune system would reflect the immuno-logic memory of the human donor. For these studies,immunodeficient NOD/SCID-IL2rγnull (NSG) mice, whichlack the γ-chain of the IL-2 receptor in addition to theNOD/SCID background, were chosen (Ishikawa et al.2005; Shultz et al. 2005). NSG mice lack high affinityreceptors for interleukin (IL)-2, IL-4, IL-7, IL-9, IL-15, andIL-21. As a result of this ineffective cytokine signaling,these mice completely lack functional T-cells, B-cells, NKcells, macrophages, and DC and allow nearly completeimmune reconstitution by engrafted human cells (Ishikawaet al. 2005; Shultz et al. 2005). We hypothesized thathumanized NSG animals, when reconstituted with bothperipheral blood lymphocytes (PBL) and with donor-matched antigen presenting DC, would allow us to examinethe process of human antigen presentation and theactivation of antigen-specific T-cells as it occurs in vivo,in the context of the immunologic repertoire of the humandonor. In these studies, we have used a replication-incompetent serotype-5 adenoviral vector (AdV) as theantigen. Human exposure to adenovirus and the develop-ment of anti-viral memory T-cell responses occurs with ahigh frequency in the adult population (Olive et al. 2001;Roth et al. 2002). As predicted, only human peripheralblood lymphocyte (huPBL)-NSG mice that were reconsti-tuted with both PBL and DC demonstrated AdV-specific T-cell proliferation and cytokine production. Furthermore, thecharacteristics of the AdV-specific response recapitulated

those of the originally implanted PBL. This model shouldprovide a rapid and highly flexible system for studying theactivation of human antigen-specific immunity and providea novel platform for vaccine development and testing, aswell as for the study of interactions between antigen-specific immunity and a variety of host challenges such asviral exposures (including HIV), tumors, and systemic drugeffects.

Materials and methods

Animals

Six- to 8-week-old NSG mice were obtained from TheJackson Laboratories (Bar Harbor, ME, USA) and housedat the UCLA Mouse/Human Chimera Core Facility underlaminar flow conditions with all food, water (acidified),caging, and bedding autoclaved before use. No antibioticswere administered. All procedures were approved by theUCLA Animal Research Committee.

Culture media and reagents

Cells were cultured in complete medium composed ofRPMI 1640 (Mediatech, Manassas, VA, USA) supple-mented with 10% heat-inactivated human AB serum(Omega Scientific, Tarzana, CA, USA), penicillin–streptomycin–fungizone (Invitrogen, Carlsbad, CA,USA), and 10 mM HEPES (Sigma Chemical, St. Louis,MO, USA). Monoclonal antibody (mAb) combinationsused for the enrichment of T-and B-cells (anti-CD14, anti-CD16, and anti-CD25) and DC (anti-CD3, anti-CD16,and anti-CD19) by negative depletion were obtained fromBD Biosciences (San Jose, CA, USA). Anti-mouseIg-conjugated immunomagnetic beads were from Invitro-gen (M450 Dynabeads; Car lsbad, CA, USA).Fluorochrome-conjugated mAbs against CD3, CD4, CD8,CD11c, CD86, and HLA-DR were obtained from BDBiosciences; anti-CD127 and anti-GITR from eBiosciences(San Diego, CA, USA); and anti-CD25 from BioLegend(San Diego, CA, USA). T-cell proliferation in vitro wasevaluated using carboxyfluorescein diacetate, succinimidylester (CFSE; Vybrant CFDA SE Cell Tracer Kit, Invitrogen-Molecular Probes, Eugene, OR, USA).

Adenoviral vectors

A replication-deficient adenoviral vector (AdV) expressinggreen fluorescent protein (GFP) and modified to express anadditional Arg-Gly-Asp peptide sequence in the HI loopregion of their fiber knob was obtained from Mizuguchi et

J Neuroimmune Pharmacol (2011) 6:148–157 149

al. (2001) (AdV-GFP). A replication-deficient adenoviralvector with a matching CMV promoter but without aninserted transgene sequence (AdV-RR5) was also used aspreviously described (Roth et al. 2002). Adenoviral vectorswere propagated in 293 cells, purified by CsCI densitycentrifugation, dialyzed, and stored in −80°C. Viral particletiters were determined spectrophotometrically (Mitterederet al. 1996).

Generation and transduction of donor-matched DC

Peripheral blood mononuclear cells (PBMC) were obtainedfrom healthy normal donors by density gradient centrifu-gation and DC prepared by culturing the adherent fractionwith 800 IU/ml granulocyte-macrophage colony stimulat-ing factor (GM-CSF; Bayer Healthcare, Seattle, WA, USA)and 63 ng/ml IL-4 (R&D Systems, Minneapolis, MN,USA) for 5 days at 37°C as previously described(Kiertscher and Roth 1996). DC were recovered on day 5and purified by negative depletion using a mAb cocktailand anti-mouse Ig-conjugated immunomagnetic beads. DCwere either used fresh or aliquoted and cryopreserved inmedium containing 20% AB serum and 10% DMSO. Freshor cryopreserved DC that had been thawed and cultured for8–24 h in complete media containing GM-CSF and IL-4were transduced with AdV as described previously (Haruiet al. 2006a, b). Briefly, 5.0×105 DC suspended in 100 μlof complete medium were exposed to a suspension of AdVin phosphate-buffered saline (PBS) containing calcium andmagnesium at a multiplicity of infection (MOI) of 4,000viral particles/cell. Cultures were incubated for 2 h at 37°Cin a humidified incubator and an additional 800 μl ofcomplete media containing GM-CSF and IL-4 was thenadded. Human MegaCD40L (200 ng/ml, Alexis Biochem-icals, Plymouth Meeting, PA, USA) and interferon gamma(IFN-γ; 100 U/ml, PeproTech, Rocky Hills, NJ, USA) wereadded into culture media 24 h later to stimulate a matureDC phenotype, and cells were incubated for an additional24 h and harvested.

Reconstitution and vaccination of NSG mice

A combination of T-and B-cells were purified fromPBMC by negative depletion using a mAb cocktail andanti-mouse Ig-conjugated immunomagnetic beads. Cellswere resuspended in PBS and administered to recipientNSG mice (107 cells/mouse) by intraperitoneal (i.p.)injection on day 0 of the reconstitution. To adapthuPBL-NSG mice as a platform for assessing humanantigen-specific T-cell activation, animals were alsoadministered an i.p. injection on day 0 of 0.5×106 DCthat had been generated from the same donor (donor-matched) (Fig. 1). Monocyte-derived DCs were adminis-

tered as either control DC (which had not been exposed toAdV in vitro) along with 1×109 or 1×1010 free AdV-GFPparticles, or as DC that had previously been transducedwith the AdV-GFP vector in vitro (DC/AdV-GFP). Forsome experiments, animals were administered a secondinjection of donor-matched DC that had been preparedfrom the same donor (either DC/AdV-GFP or control DCin combination with free viral particles) 7 days later usingthe same methods.

Cell recovery and analysis of human cells by flowcytometry

Single-cell suspensions were prepared from the spleen bymechanical disaggregation and from the peritoneal cavity bylavage followed by depletion of red blood cells by hypotonicshock. Cells were stained with fluorochrome-conjugatedmAbs and analyzed by flow cytometry (FacsCalibur,Becton-Dickinson, San Jose, CA, USA) in combination withFCS Express analysis software (DeNovo Software, LosAngeles, CA, USA). Human cells were defined by expressionof human CD45 and subsets identified by specific markerexpression for T-cell subsets (CD3+/CD4+ or CD3+/CD8+alone or in combination with GITR, CD127, and/or CD25),B-cells (CD20), NK cells (CD3-/CD56+), NK-T-like cells(CD3+/CD56+), or DC (CD11c, CD40, CD86, and/or HLA-DR).

Detection of antigen-specific T-cell responses

Recovered spleen cells were challenged in vitro withdonor-matched control DC or with DC/AdV-RR5 todetect responses to adenoviral capsid antigens. Antigen-specific responses were assessed by intracellular cytokinestimulation (ICCS), proliferation, or cytokine secretionassays. For ICCS, 2×106 spleen cells were stimulatedwith either control DC or DC/AdV-RR5 at a ratio of 1:10(DC/spleen cells ratio) in complete media in the presenceof 3 µg/ml anti-CD28 Ab (BD Biosciences) and 1 ng/mlIL-12 (PeproTech) for 5 h at 37°C. Brefeldin A (Sigma-Aldrich, St. Louis, MO, USA) was added at 10 µg/ml afterthe first hour. At the end of the culture period, cells werefixed and cryopreserved. Intracellular staining for IFN-γand tumor necrosis factor alpha (TNF-α) was then carriedout according to the manufacturer’s protocol (BDBiosciences) in combination with counterstaining usingfluorochrome-conjugated anti-CD3, anti-CD4, and anti-CD8 antibodies. A minimum of 150,000 CD3+ cell eventswere analyzed by multi-color flow cytometry. For prolif-eration assays, spleen cells were first labeled with CFSE(0.5 µM) and then 2×105 cells/well co-cultured witheither control DC or DC/AdV-RR5 at a ratio of 1:20 (DC/spleen cells ratio) in the presence of low dose IL-2 (10 U/

150 J Neuroimmune Pharmacol (2011) 6:148–157

ml) and IL-7 (1 ng/ml). Cells were harvested 5 days later,counterstained with fluorescent-conjugated mAb to indi-cated markers and analyzed by flow cytometry. Antigen-specific cytokine secretion was assessed by collectingsupernatants from the same wells used for proliferationassays and the levels of human IFN-γ and TNF-αmeasured by SearchLight multiplex assay (Thermo Scien-tific, Rockford, IL, USA).

In vivo cytokine production

Serum samples were collected from animals at the timeof spleen recovery and assayed for the presence of

human IFN-γ and TNF-α to determine whetherreconstitution with control DC versus DC/AdV-GFPresulted in detectable differences in systemic cytokinelevels, reflecting the in vivo activation of antigen-specific T-cells.

Statistical analysis

Comparisons between groups were performed using Stu-dent’s t test, with p<0.05 accepted as a significantdifference between groups. All experiments were per-formed with replicates and results from at least threeseparate experiments were pooled for statistical analysis.

107 huPBL

NSG Mouse

+ DC/AdV-GFPor

DC + AdV-GFP

0 7 14 21

Harvest spleen, peritoneal cellsand serum

huPBL-NSG

a

c Dendritic Cells (DC/AdV-GFP)

FSC

SS

C

CD86HLA-DR

# of

eve

nts

GFP

71.3%

CD11c

b huPBL

6.80%93.2%19.9%80.1%

CD20CD56CD8S

SC

CD

3

CD

4

1.52%46.5%

33.6%

Fig. 1 Reconstitution of huPBL-NSG mice with both PBL and donor-matched dendritic cells. a Six- to 8-week-old NSG mice wereimplanted with 107 purified human T-and B-cells (PBL) by i.p.injection. In order to restore antigen-presenting activity, animals werealso administered donor-matched monocyte-derived DC alone or incombination with adenoviral antigens (as either DC/AdV-GFP orcontrol DC + free AdV-GFP particles) by i.p. injection. In someexperiments, animals received a single administration of 5×105 DCand in other experiments they received a second dose of DC 7 dayslater. Spleens, peritoneal cells, and serum were recovered on day 14 or21 for assessment of antigen-specific immune responses. b Purifiedhuman PBL used for implantation into NSG mouse were composed of

subsets of T-cells (CD3+/CD4+, CD3+/CD8+, and CD3+/CD56+) andB-cells. c Monocyte-derived DC generated from the same donorexpress high levels of antigen and maturation markers. DC weretransduced with AdV-GFP at an MOI of 4,000 viral particles/cell andwere stained with antibodies against DC phenotype/maturationmarkers including CD11c, HLA-DR, and CD86 48 h later. Flowcytometry demonstrated a uniform population of DC by forwardscatter (FSC) and side scatter (SSC). The percentage of GFP-positivecells was used to measure the percentage of DC expressing AdVantigens after transduction. The fluorescent profile of unstained DC ispresented in gray and fluorescent staining for specific markers aredepicted by black histogram curves. Representative experiment, n=5

J Neuroimmune Pharmacol (2011) 6:148–157 151

Results

Reconstitution of huPBL-NSG mice with both PBLand donor-matched dendritic cells

To adapt huPBL-NSG mice as a platform for assessinghuman antigen-specific T-cell activation, animals wereimplanted with an i.p. injection of 107 human PBL andthen supplemented with 0.5–1.0×106 DC that had beengenerated from the same donor (donor-matched) as detailedin Fig. 1. Donor PBL were composed primarily of CD3+ T-cells with a normal complement of CD4 and CD8 subsets, apopulation of CD3+/CD56+ cells, and CD20+ B-cells(Fig. 1b). Culture of adherent PBMC from the same donorwith GM-CSF and IL-4, followed by immunomagneticdepletion of contaminating lymphocytes, resulted in arelatively pure and homogenous population of DC(Fig. 1c). DC prepared in this manner were highlysusceptible to transduction by AdV and expressed highlevels of the characteristic CD11c marker, MHC molecules(HLA-DR), and costimulatory molecules (such as CD86)required for efficient T-cell activation.

Immune reconstitution of NSG mice with these twopopulations was associated with a marked migration to, andexpansion of, human cells populating the spleen over thefirst 2–3 weeks. Splenic repopulation was associated with aconcomitant decrease in the number of intraperitoneal cells(Fig. 2a). By the end of 3 weeks, the average yield ofhuman cells from the spleen was 19.4±6.7×106 CD45+cells and the number of residual CD45+ cells in theperitoneum had decreased to 1.9±0.6×106. After 14 days,78.1±3.0% (n=3) of recovered spleen cells expressedhuman CD45 and the percentage of human cells remainedrelatively constant through 21 days (Fig. 2b).

Recovered spleen cells contained a distribution oflymphocyte subsets similar to that of normal human blood

including CD3+/CD4+, CD3+/CD8+, and CD3+/CD56+ T-cells along with a small population of B-cells (CD20+)(Fig. 3a). The CD4/CD8 ratio averaged 1.63:1 (n=15) andthere was a direct relationship between the frequency ofCD3+/CD56+ T-cells in the PBL used for implantation andthose covered from the spleen 2–3 weeks later (r2=0.85,p<0.01). There was little evidence for T-cell activation asmeasured by the expression of either CD25 or CD69 (datanot shown). The i.p. administration of mature monocyte-derived DC into huPBL-NSG mice also resulted in theirmigration and repopulation of the spleen by the end of2 weeks. Similar to the mature phenotype of these cells atthe time of administration, the majority of recoveredCD11c+ splenic DC expressed high levels of HLA-DRand CD86 (>70%) (Fig. 3b). huPBL-NSG mice recon-stituted in this manner exhibited no physical signs of graftversus host disease.

Activation of antigen-specific T-cells in vivo

The capacity for human DC derived from the same donorto interact with and stimulate antigen-specific T-cells invivo in huPBL-NSG animals was evaluated under twodifferent conditions. In the first approach, animals wereadministered DC in combination with free AdV-GFPparticles. In the second approach, animals were adminis-tered DC that had been loaded 48 h earlier with AdVcapsid antigens in vitro. In both settings, the stimulationof memory T-cell responses to AdV capsid antigens wasassessed by a standard 5-h ICCS assay in which non-specific T-cell activation was assessed by challengingrecovered cells with control DC (containing no antigen)while AdV-specific T-cell immunization was assessed bychallenging with DC that had been loaded with theAdV/RR5 vector. Figure 4 demonstrates the results froma typical experiment. The baseline frequency of AdV-

SS

C-H

81.2% 77.3%

Spleen Peritoneumb

CD45 CD450

15

30

14 days 21 days

CD

45+

(x1

06 )

a Spleen Peritoneum

Fig. 2 Engraftment of NSG mice with human PBL. Six- to 8-week-old NSG mice were implanted with 107 purified human T-and B-cells(PBL) by i.p. injection. Spleen and peritoneal cells were recoveredafter 14 or 21 days, counted, and stained with anti-human CD45 mAbfor analysis by flow cytometry. a The increase in human CD45+leukocytes within the spleen from 14 to 21 days was associated with a

concomitant decrease in peritoneal cell number, suggesting acombined redistribution and expansion of human cells following i.p.implantation. b Representative dot-plots demonstrating the percentageof recovered cells from the spleen and peritoneum that expressedhuman CD45 at day 14. Average percentage of human CD45+ cells inthe spleen at day 14 was 78.1±3.0%, n=3

152 J Neuroimmune Pharmacol (2011) 6:148–157

specific T-cells capable of producing IFN-γ and TNF-αbefore implantation was measured at 0.10% of CD4+ T-cells (all reported frequencies adjusted for the response tocontrol DC) and 0.03% of CD8+ T-cells (Fig. 4a). When

PBL were administered to NSG mice in the absence ofDC, the frequency of AdV-specific CD4+ T-cells recov-ered from day 21 spleen cells was always less than 0.02%(data not shown). However, huPBL-NSG mice that were

a Lymphocyte Subsets

4.47%

0.00%

95.5%

0.00%

30.7%

0.50%

68.7%

0.17%

CD20CD56CD8

CD

45

CD

3

CD

4

7.74%

1.29%

52.7%

38.2%

b Dendritic Cells (CD11c+)

CD86

HLA

-DR

78%

CD40

HLA

-DR

72%

Fig. 3 Lymphocyte and DC phenotypes reconstituting the spleens ofhuPBL-NSG mice. a Six- to 8-week-old NSG mice were implantedwith 107 purified human PBL at day 0 in combination with 5×105

donor-matched monocyte-derived DC by i.p injection at day 0 andday 7. Spleens were recovered 14 days later and cells expressinghuman CD45 analyzed by flow cytometry. CD3+ T-cells exhibited anormal distribution of CD4+ and CD8+ subsets (left); a smallpopulation of B-cells was identified by their expression of CD20(right); and NK-T-like cells were identified by their expression of

CD56 alone or in combination with CD3, respectively (center).Representative experiment, n=15. b Six- to 8-week-old NSG micewere implanted with 107 purified human PBL at day 0 and 5×105

donor-matched monocyte-derived DC by i.p injection at day 7.Spleens were recovered 14 days later and DC were identified bygating on the population of cells expressing CD11c and analyzing theCD11c+ subset for their expression of both HLA-DR and CD40 (left,gated for CD45+/CD11c+ cells) and both HLA-DR and CD86 (right,gated for CD45+/CD11c+ cells). Representative experiment, n=5

a

0.03%

0.10%

IFN-γ/TNF-α

In vitro DC/AdV-GFP

CD

4

0.03%

0.14%

IFN-γ/TNF-α

In vivo DC + AdV-GFP

CD

4

b

0.02%

0.39%

IFN-γ/TNF-α

In vivo DC/AdV-GFP

CD

4

c

Fig. 4 Donor-matched DC loaded with AdV antigens stimulateantigen-specific T-cell responses in vitro and in vivo in huPBL-NSGmice. Adherent PBMC were cultured with GM-CSF and IL-4 for5 days, transduced with 4,000 MOI AdV-GFP (DC/AdV-GFP), with4,000 MOI AdV-RR5 (DC/AdV-RR5), or cultured in transductionmedium alone (control DC), and then matured with MegaCD40L andIFN-γ. a The capacity to stimulate antigen-specific T-cells in vitrowas assessed by co-culturing DC/AdV-GFP with purified donor-matched T-cells (DC/T ratio=1:20) for 7 days followed by challengewith either control DC or DC/AdV-RR5 to detect AdV-specific T-cellfrequency in a standard 5-h ICCS assay. Frequencies of antigen-specific CD4+ and CD8+ T-cells expressing IFN-γ and/or TNF-α are

noted, respectively (all frequencies are adjusted for the backgroundresponse to control DC in the absence of antigen). Representativeexperiment, n=3. b, c The capacity for DC to stimulate antigen-specific T-cells in vivo in huPBL-NSG mice was similarly assessed byimmunizing animals with either b 5×105 DC/AdV-GFP (on day 7) orwith c 5×105 control DC (on day 7) in combination with 2×109 freeAdV-GFP particles. Spleen cells were harvested 14 days later andchallenged in vitro with either control DC or DC/AdV-RR5 to detectAdV-specific T-cell frequency in a standard 5-h ICCS assay.Frequencies of antigen-specific CD4+ and CD8+ T-cells expressingIFN-γ and/or TNF-α are noted, respectively. Representative experi-ment, n=4

J Neuroimmune Pharmacol (2011) 6:148–157 153

immunized in vivo with the combination of DC and freeAdV particles exhibited an AdV-specific CD4+ T-cellfrequency of 0.14% (Fig. 4b). Furthermore, when huPBL-NSG mice were immunized in vivo with DC/AdV-GFP,the frequency of AdV-specific CD4+ T-cells increasedmarkedly to 0.39% (Fig. 4c). In replicate experiments, theAdV-specific CD4+ T-cell responses in animals that wereimmunized with DC/AdV-GFP were always greater thanthe responses in animals that received control DC incombination with free AdV-GFP. When administering freeAdV-GFP, the response to 1010 viral particles (shown inFig. 4b) was generally greater than the response to109 viral particles (data not shown).

Antigen-specific cytokine production induced by DC-basedimmunization

Sera from immunized huPBL-NSG mice were tested for thepresence of human IFN-γ and TNF-α (Fig. 5). Comparedto serum from control animals that were immunized withDC alone (no AdV exposure), serum from animalsimmunized with DC/AdV-GFP demonstrated consistentlyincreased levels of both cytokines consistent with theresponses measured in the ICCS assay.

Antigen-specific T-cells generated by immunizationof huPBL-NSG mice respond to a secondaryin vitro challenge

To further evaluate the function of antigen-specific T-cellsthat were activated in vivo, spleen cells were recoveredfrom huPBL-NSG mice that were immunized withDC/AdV-GFP. The cells were labeled with CFSE and re-challenged in vitro for 5 days with either control DC(containing no antigens) or with DC/AdV-RR5 (Fig. 6).Although some background proliferation was observed inresponse to control DC, there was a clear antigen-specificincrease in the percentage of proliferating CD4+ andCD4− T-cells in response to challenge with DC/AdV-RR5

(Fig. 6a). Further phenotyping of the proliferating cellsrevealed that they expressed CD45RO, a memory cellmarker, and that a significant percentage expressed CD25,an activation marker, in the absence of GITR, suggestingthat they were not suppressor cells (Fig. 6b). Consistentwith the expansion of effector/memory T-cells, supernatantcollected from these co-cultures demonstrated antigen-specific induction of IFN-γ and TNF-α (Fig. 6c).

0

50

100

150

200

250

DC DC/AdV-GFP

IFN

- (

pg/m

l in

seru

m)

0

2

4

6

8

10

12

DC DC/AdV-GFP

TN

F-

(pg/

ml i

n se

rum

)

*

IFN

- (

pg/m

l in

seru

m)

*

Fig. 5 In vivo cytokine response to DC/AdV-GFP. Sera from huPBL-NSG mice were collected 14 days after immunization with eithercontrol DC or DC/AdV-GFP and assayed for human IFN-γ and TNF-α by SearchLight multiplex assay. Representative experiment, n=3.†p<0.10; *p≤0.05

c In vitro cytokine productionIF

N-γ

(pg/

ml)

0

50

100

150

200

DC DC/AdV-RR50

10

20

30

40

DC DC/AdV-RR5T

NF

-α (p

g/m

l)

a In vitro stimulation of T-cells from huPBL-NSG mice previously immunized in vivo with DC/AdV-GFP

Control DC DC/AdV-RR5

CFSE

CD

4

CFSE

CD

4

18.6%

11.3%

11.9%

6.6%

0.36%

1.39%

5.33%

92.9%

CD45RAC

D45

RO

2.31%

0.62%23.5%

74.6%

GITR

CD

25

b Phenotype of proliferating T-cells (DC/AdV-RR5 group)

* *

Fig. 6 T-cells recovered from immunized huPBL-NSG mice respondto antigen-specific stimulation in vitro. Spleen cells recovered fromhuPBL-NSG mice that were previously immunized with DC/AdV-GFP in vivo were labeled with CFSE and co-cultured with eithercontrol DC or DC/AdV-RR5 (DC/T ratio=1:20) for 5 days in vitro. aCells were stained with anti-CD3 and anti-CD4 and the CD3+ T-cellsanalyzed for adenoviral-specific proliferative responses by flowcytometry. Representative experiment, n=5. b The sub-population ofcells proliferating in response to stimulation with DC/AdV-RR5(CFSE-dim) were further analyzed for their expression of effector/memory and regulatory T-cell markers. c Supernatants from the samewells were harvested and assessed for the presence of human IFN-γand TNF-α by SearchLight multiplex assay. Representative experi-ment, n=3. *p<0.01

154 J Neuroimmune Pharmacol (2011) 6:148–157

Discussion

Fundamental differences between the mouse and humanimmune systems, differences in genetic diversity, as well asinherent differences in the susceptibility and exposure tohuman disease, often limit the translational potential of mousemodels for developing effective treatments for human disease.Avariety of “humanized” animal models have been developedto address these issues by either expressing human genes/traitsin animals (Pascolo et al. 1997; Mangalam et al. 2008;Conforti et al. 2009) or by engrafting human immune cellsinto immunodeficient mice (Mosier et al. 1988; Ito et al.2002; Ishikawa et al. 2005). However, one significantlimitation has been the study of adaptive immunity. ThehuPBL-NSG model provides a unique opportunity to studythe interaction of various immune challenges with humanimmune cells (Kumar et al. 2008; Strowig et al. 2009; Kinget al. 2009). We herein demonstrate that the addition ofdonor-matched monocyte-derived DC to the reconstitution ofhuPBL-NSG animals enabled us to rapidly assess andstimulate the donor’s antigen-specific memory in vivo, inthe intact animal, as well as ex vivo, after recovery of spleencells from immunized animals.

Our results confirm that rapid reconstitution of thespleen and peripheral blood occurs following a single i.p.injection of human PBL. Within a few weeks, 15–25million human cells can be recovered from the spleen withthe CD3+/CD4+, CD3+/CD8+, CD3+/CD56+, and CD20+populations all well represented in appropriate proportions.Regardless of whether administered only once (at day 0 or7) or twice (at both day 0 and 7), implanted DC weredetected at low levels in the spleen and expressed thephenotypic markers of functional antigen-presenting cells.Importantly, reconstitution with DC did not lead to physicalsigns of GVHD or to changes in the level of acuteactivation markers (CD25 and CD69; data not shown).However, reconstitution with DC did enable the huPBL-NSG animals to respond to antigen challenge. Adenoviruswas chosen as the target antigen for this work due to itswell-characterized immunogenicity in humans and thepresence of highly sensitive assays for assessing AdV-specific T-cell responses (Roth et al. 2002). Similar to theresponse pattern observed in vivo when humans areimmunized with AdV (Gahéry-Ségard et al. 1997), fullyreconstituted huPBL-NSG animals rapidly expanded thepopulation of AdV-specific T-cells following exposure toAdV. In the absence of an antigen challenge, essentially nodetectable AdV-specific T-cells were detected in freshlyharvested splenocytes. However, following active immuni-zation with DC/AdV-GFP, there was a dramatic upregula-tion of AdV-specific T-cells that far exceeded the leveldetected when the original donor PBL were stimulated invitro for a week with IL-2, IL-7, and DC/AdV-GFP. As

might be expected when healthy patients are immunizedwith a vaccine, immunization of the intact animal was moreeffective than standard in vitro stimulation. While a cell-based vaccination (DC/AdV-GFP) elicited the greatest invivo response, administration of free AdV-GFP vector alsostimulated an AdV-specific T-cell response. In this case, themagnitude of the response appeared to increase with thequantity of antigen administered.

In addition to the in vivo expansion of antigen-specificT-cells, immunization of huPBL-NSG animals with DC/AdV-GFP induced the systemic release of IFN-γ and TNF-α into the serum. The detection of these circulatingcytokines was a robust indicator of the AdV-specific T-cell response and consistent with reports that IFN-γ andTNF-α represent the predominant T-cell cytokines pro-duced upon AdV challenge (Hutnick et al. 2010). It isnoteworthy that resting cytokine levels for IFN-γ and TNF-α were low in unchallenged animals, arguing against thepresence of spontaneous responses to endogenous mouseantigens.

One advantage of the huPBL-NSG mouse compared toits NOD/SCID precursors is the robust repopulation of thespleen with human cells. Harvesting spleens from theseanimals provides an abundant source of T-cells for in vitroanalysis. As occurs when human T-cells are recovered fromthe blood of healthy subjects, the T-cells recovered from thespleens of immunized animals were able to undergo furtherexpansion and testing when stimulated with AdV foranother 5–7 days in vitro. Labeling recovered spleen cellswith CFSE allowed us to track the proliferative response toa secondary in vitro challenge with DC/AdV-RR5 and tocharacterize the phenotype of responding cells. Interesting-ly, this was the one situation in which there was asignificant background response when the cells werestimulated with control DC that had not been loaded withAdV. As the co-cultures contained up to 30% mouse cellsrecovered from the spleen, it seems likely that uptake andprocessing of these cells by the added DC lead to an in vitroanti-mouse response. However, despite this higher back-ground proliferation, challenging in vitro with DC/AdV-RR5 still led to a significantly greater response than didchallenging with control DC alone. T-cells proliferating inresponse to DC/AdV-RR5 were mostly CD45RO+ cells andapproximately 25% of them expressed moderate levels ofCD25 without expression of GITR, suggesting the presenceof memory/effector T-cells, but not T-regulatory cells. Thiswould be expected given our use of mature DC toimmunize the animals in vivo (Roth et al. 2002; Harui etal. 2004). As detected in serum collected from immunizedanimals, IFN-γ and TNF-α were secreted into the culturesupernatant during the in vitro challenge.

In summary, these findings support our hypothesis thatreconstitution of NSG mice with a combination of human

J Neuroimmune Pharmacol (2011) 6:148–157 155

PBL and mature donor-matched human DC results in ahuPBL-NSG model in which adaptive T-cell responses canbe stimulated and measured in the intact animal. There areseveral potential advantages of this model over others thathave been reported in the literature. Gorantla et al. (2005)reported that NOD/SCID animals can be treated withasialo-GM1 and anti-CD122 antibodies to aid humanengraftment and then vaccinated with antigen-loaded DC.While providing an important proof of concept, the use ofNOD/SCID animals for this purpose has been replaced bynewer animal strains due to the large number of cellsrequired and the very limited engraftment that is obtained(Shultz et al. 2000; Shultz et al. 2003). Conventionalengraftment of NSG mice with CD34+ stem cells is limitedby the failure of normal T-cell development and the lack ofpre-existing immune memory (Ishikawa et al. 2005). Yu etal. (2008) overcame these limitations by transplantingNOD/SCID β2 microglobulin−/− mice with peripheralblood CD34+ stem cells (3×106/animal) to promote DCand B-cell repopulation followed by supplementation4–8 weeks later with donor-matched PBL as the source ofT-cell responders. While requiring a leukopheresis to obtainsufficient CD34 donor cells, whole-body irradiation incombination with a 10-day course of human FLT3 ligand topromote engraftment, and then a protracted period forreconstitution, the resulting humanized animals respondedwell to both live influenza virus and inactivated influenzavaccine. As in our model, they observed a significant invivo expansion of memory/effector T-cells with definedantigen specificity and cytokine production. Also similarto our work, they demonstrated that DC were crucial forthe generation of an antigen-specific response. Whilebased on similar principals, our model allows reconstitu-tion from a single peripheral blood draw, and animals areready for immunologic testing in as little as 2 weekswithout further manipulation. In addition, these animalsare stable for at least 6 weeks after implantation of humanPBL. The large number of human cells that can berecovered from the spleen facilitates the extensive in vitroanalysis that is often required when mapping out T-cellresponses. Altogether, this model provides a highlyadaptable platform to study the interaction of antigenpresentation and antigen-specific T-cell responses in vivo.Immunologic responses to a variety of viruses can likelybe studied using these approaches, new vaccines testedwithout the need to administer them to human subjects,and a variety of drugs and immune modulators evaluatedfor their impact on various stages of antigen presentationand T-cell function.

Conflict of interest statement The authors do not have financialdisclosure, personal relationship, or known conflict of interest relatedto this work.

Open Access This article is distributed under the terms of theCreative Commons Attribution Noncommercial License which per-mits any noncommercial use, distribution, and reproduction in anymedium, provided the original author(s) and source are credited.

References

Conforti A, Peruzzi D, Giannetti P, Biondo A, Ciliberto G, La MonicaN, Aurisicchio L (2009) A novel mouse model for evaluation andprediction of HLA-A2-restricted CEA cancer vaccine responses.J Immunother 32(7):744–54

Gahéry-Ségard H, Molinier-Frenkel V, Le Boulaire C, Saulnier P,Opolon P, Lengagne R, Gautier E, Le Cesne A, Zitvogel L, VenetA, Schatz C, Courtney M, Le Chevalier T, Tursz T, Guillet JG,Farace F (1997) Phase I trial of recombinant adenovirus genetransfer in lung cancer. Longitudinal study of the immuneresponses to transgene and viral products. J Clin Invest 100(9):2218–26

Gorantla S, Santos K, Meyer V, Dewhurst S, Bowers WJ, Federoff HJ,Gendelman HE, Poluektova L (2005) Human dendritic cellstransduced with herpes simplex virus amplicons encoding humanimmunodeficiency virus type 1 (HIV-1) gp120 elicit adaptiveimmune responses from human cells engrafted into NOD/SCIDmice and confer partial protection against HIV-1 challenge. JVirol 79(4):2124–32

Greiner DL, Hesselton RA, Shultz LD (1998) SCID mouse models ofhuman stem cell engraftment. Stem Cells 16(3):166–77

Harui A, Roth MD, Kiertscher SM, Mitani K, Basak SK (2004)Vaccination with helper-dependent adenovirus enhances thegeneration of transgene-specific CTL. Gene Ther 11(22):1617–26

Harui A, Roth MD, Vira D, Sanghvi M, Mizuguchi H, Basak SK(2006a) Adenoviral-encoded antigens are presented efficiently bya subset of dendritic cells expressing high levels of alpha(v)beta3integrins. J Leukoc Biol 79(6):1271–8

Harui A, Roth MD, Sanghvi M, Vira D, Mizuguchi H, Basak SK(2006b) Centrifugation enhances integrin-mediated transductionof dendritic cells by conventional and RGD-modified adenoviralvectors. J Immunol Methods 312(1–2):94–104

Hesselton RM, Greiner DL, Mordes JP, Rajan TV, Sullivan JL, ShultzLD (1995) High levels of human peripheral blood mononuclearcell engraftment and enhanced susceptibility to human immuno-deficiency virus type 1 infection in NOD/LtSz-scid/scid mice. JInfect Dis 172(4):974–82

Hutnick NA, Carnathan D, Demers K, Makedonas G, Ertl HC, BettsMR (2010) Adenovirus-specific human T-cells are pervasive,polyfunctional, and cross-reactive. Vaccine 28(8):1932–41

Ishikawa F, Yasukawa M, Lyons B, Yoshida S, Miyamoto T,Yoshimoto G, Watanabe T, Akashi K, Shultz LD, Harada M(2005) Development of functional human blood and immunesystems in NOD/SCID/IL2 receptor gamma chain (null) mice.Blood 106(5):1565–73

Ito M, Hiramatsu H, Kobayashi K, Suzue K, Kawahata M, Hioki K,Ueyama Y, Koyanagi Y, Sugamura K, Tsuji K, Heike T,Nakahata T (2002) NOD/SCID/gamma(c)(null) mouse: anexcellent recipient mouse model for engraftment of human cells.Blood 100(9):3175–82

Ju X, Clark G, Hart DN (2010) Review of human DC subtypes. MethMol Biol 595:3–20

Kiertscher SM, Roth MD (1996) Human CD14+ leukocytes acquirethe phenotype and function of antigen-presenting dendritic cellswhen cultured in GM-CSF and IL-4. J Leukoc Biol 59(2):208–18

156 J Neuroimmune Pharmacol (2011) 6:148–157

KingMA, Covassin L, BrehmMA, RackiW, Pearson T, Leif J, Laning J,Fodor W, Foreman O, Burzenski L, Chase TH, Gott B, Rossini AA,Bortell R, Shultz LD, Greiner DL (2009) Human peripheral bloodleucocyte non-obese diabetic-severe combined immunodeficiencyinterleukin-2 receptor gamma chain gene mouse model of xenoge-neic graft-versus-host-like disease and the role of host majorhistocompatibility complex. Clin Exp Immunol 157(1):104–18

Kumar P, Ban HS, Kim SS, Wu H, Pearson T, Greiner DL, Laouar A,Yao J, Haridas V, Habiro K, Yang YG, Jeong JH, Lee KY, KimYH, Kim SW, Peipp M, Fey GH, Manjunath N, Shultz LD, LeeSK, Shankar P (2008) T cell-specific siRNA delivery suppressesHIV-1 infection in humanized mice. Cell 134(4):577–86

Lapidot T, Pflumio F, Doedens M, Murdoch B, Williams DE, Dick JE(1992) Cytokine stimulation of multilineage hematopoiesis fromimmature human cells engrafted in SCID mice. Science 255(5048):1137–41

Mangalam AK, Rajagopalan G, Taneja V, David CS (2008) HLA classII transgenic mice mimic human inflammatory diseases. AdvImmunol 97:65–147

May KF Jr, Roychowdhury S, Bhatt D, Kocak E, Bai XF, Liu JQ,Ferketich AK, Martin EW Jr, Caligiuri MA, Zheng P, Liu Y(2005) Anti-human CTLA-4 monoclonal antibody promotes T-cell expansion and immunity in a hu-PBL-SCID model: a newmethod for preclinical screening of costimulatory monoclonalantibodies. Blood 105(3):1114–20

Mestas J, Hughes CC (2004) Of mice and not men: differences betweenmouse and human immunology. J Immunol 172:2731–2738

Mittereder N, March KL, Trapnell BC (1996) Evaluation of theconcentration and bioactivity of adenovirus vectors for genetherapy. J Virol 70(11):7498–509

Mizuguchi H, Koizumi N, Hosono T, Utoguchi N, Watanabe Y, KayMA, Hayakawa T (2001) A simplified system for constructingrecombinant adenoviral vectors containing heterologous peptidesin the HI loop of their fiber knob. Gene Ther 8(9):730–5

Mosier DE, Gulizia RJ, Baird SM, Wilson DB (1988) Transfer of afunctional human immune system to mice with severe combinedimmunodeficiency. Nature 335:256–259

Olive M, Eisenlohr LC, Flomenberg P (2001) Quantitative analysis ofadenovirus-specific CD4+ T-cell responses from healthy adults.Viral Immunol 14(4):403–13

Pascolo S, Bervas N, Ure JM, Smith AG, Lemonnier FA, Pérarnau B(1997) HLA-A2.1-restricted education and cytolytic activity of

CD8(+) T lymphocytes from beta2 microglobulin (beta2 m)HLA-A2.1 monochain transgenic H-2Db beta2m double knock-out mice. J Exp Med 185(12):2043–51

Pearson T, Greiner DL, Shultz LD (2008) Humanized SCID mousemodels for biomedical research. Curr Top Microbiol Immunol324:25–51

Roth MD, Cheng Q, Harui A, Basak SK, Mitani K, Low TA,Kiertscher SM (2002) Helper-dependent adenoviral vectorsefficiently express transgenes in human dendritic cells but stillstimulate antiviral immune responses. J Immunol 169(8):4651–6

Shultz LD, Lang PA, Christianson SW, Gott B, Lyons B, UmedaS, Leiter E, Hesselton R, Wagar EJ, Leif JH, Kollet O,Lapidot T, Greiner DL (2000) NOD/LtSz-Rag1null mice: animmunodeficient and radioresistant model for engraftment ofhuman hematolymphoid cells, HIV infection, and adoptivetransfer of NOD mouse diabetogenic T-cells. J Immunol164:2496–2507

Shultz LD, Banuelos S, Lyons B, Samuels R, Burzenski L, Gott B,Lang P, Leif J, Appel M, Rossini A, Greiner DL (2003)NOD/LtSz-Rag1nullPfpnull mice: a new model system toincrease levels of human peripheral leukocyte and hematopoieticstem cell engraftment. Transplantation 76:1036–1042

Shultz LD, Lyons BL, Burzenski LM, Gott B, Chen X, Chaleff S, KotbM, Gillies SD, King M, Mangada J, Greiner DL, Handgretinger R(2005) Human lymphoid and myeloid cell development in NOD/LtSz-scid IL2R gamma null mice engrafted with mobilized humanhemopoietic stem cells. J Immunol 174(10):6477–89

Shultz LD, Ishikawa F, Greiner DL (2007) Humanized mice intranslational biomedical research. Nat Rev Immunol 7(2):118–30

Strowig T, Gurer C, Ploss A, Liu YF, Arrey F, Sashihara J, Koo G,Rice CM, Young JW, Chadburn A, Cohen JI, Münz C (2009)Priming of protective T-cell responses against virus-inducedtumors in mice with human immune system components. J ExpMed 206(6):1423–34

Yu CI, Gallegos M, Marches F, Zurawski G, Ramilo O, García-SastreA, Banchereau J, Palucka AK (2008) Broad influenza-specificCD8+ T-cell responses in humanized mice vaccinated withinfluenza virus vaccines. Blood 112(9):3671–8

Zaiss AK, Machado HB, Herschman HR (2009) The influence ofinnate and pre-existing immunity on adenovirus therapy. J CellBiochem 108(4):778–90

J Neuroimmune Pharmacol (2011) 6:148–157 157